While at Rutgers, I worked with Mike Fehling on a computer system designed to learn by reflection. Called Schemer, it evolved into a general-purpose software environment for constructing knowledge-based systems.

Building on Kant’s critique of knowledge, especially his theory of synthetic a priori knowledge, the system used schemas as the basic elements. A dozen or so years later, the schema notion appeared again in the work of the Center for the Study of Reading. It was used there as a key to understanding reading comprehension.

The emphasis on the nature of knowledge and how a computer might in some sense possess knowledge led Mike and me to Schopenhauer’s The World as Will and Representation and Wittgenstein’s ideas on representation and language games. We also co-taught a senior capstone class on Models of Thought.

Work with another psychologist, Charles Schmidt, led to theories of social action, causation, and attribution theory (Fritz Heider). This was used in the Believer system for natural language understanding.

I also participated in a reading/discussion group on structuralism, focusing especially on Jean Piaget, but also Claude Lévi-Strauss, Noam Chomsky, Plato, etc.